IRC Logs for #crux-devel Tuesday, 2014-03-18

jaegerbeen out of town for a family emergency, just got back00:24
*** frinnst has quit IRC00:27
*** frinnst has joined #crux-devel00:27
*** frinnst has quit IRC00:27
*** frinnst has joined #crux-devel00:27
prologicwb :)00:28
prologicfrinnst, jaeger: (from the ml) re "resources to mirror package sources"00:28
prologicwhat kind of resources would be required to sport this?00:28
jaegerI started working on a script a while back to automatically create a source mirror and keep it updated, would probably just need something like that plus the bandwidth and uptime. My plan was to test it on my VPS for a while but more important things came up00:32
prologicFeel like throwing your script my way to have a look?00:36
prologicThe question for me (if I were to help provide such a service) would be:00:36
prologica) How much traffic will this consume?00:36
prologicb) How much storage is required?00:36
jaegerIt wasn't nearly finished, would need much more work00:36
prologicBecause those two things cost $$$ :)00:36
jaegerno idea on a, b would be easy to figure out once everything is downloaded00:36
prologicThe reason I'm bringing this up (again)00:37
prologicIs I plan to move my Codeo based server to Hetzne.de00:37
prologicwhich has a much better offer for less money00:37
prologicSo I may be able to satisfy my own needs00:37
prologicplus a+b for CRUX00:37
prologicas a source mirror service00:37
jaegerCurrently I want to get a 3.1 test ISO made, will add the source mirror thing back to my list to check out00:39
prologicMind if I have a go at writing firstly such a tool?00:40
jaegerfeel free00:40
prologicwhat features should it have?00:40
prologicincluding ones you never finished00:41
jaegerMine was going to be very simple, just maintain current sources (so periodically download new versions after Pkgfile updates) and clean up old sources after some amount of time00:41
jaegerno rocket surgery needed00:41
prologickk that sounds fairly straight forward00:42
prologiccould even run (or part thereof) as a git hook00:42
jaegerStill have no idea what to do about the EFI problem with 3.1 ISO booting. It isn't the loader. elilo and grub2 produce the same results but both work later to boot an installation done in legacy mode00:47
prologicSo UEFI booting is broken atm?01:01
jaegeryes. everything else seems ok01:05
prologicdamn :/ that'll suck for anyone with UEFI hw01:08
prologicand can't turn it off (are there any such devices?)01:08
jaegerYou can still do an install in legacy mode and boot UEFI later, just doesn't work on the ISO01:08
jaegerthe only UEFI-only devices I know about are macs and they can boot in legacy mode via bootcamp01:08
prologicfair enough01:12
prologicso then no, there are no such restricted devices01:12
prologicthank goodness01:12
prologicI'm sure if DRM supporters if they have their way will force this one day soon01:12
jaegeryou can use secure boot with linux already, just have to boot with the right signed binary and chainload to something else01:13
jaegerfor what that's worth01:13
prologicit's worth nothing :)01:15
jaegerIt's worth something if you don't want to disable secure boot01:15
*** jue has quit IRC01:16
*** jue has joined #crux-devel01:25
*** nomius has quit IRC01:54
*** nomius has joined #crux-devel02:01
*** c0x` has joined #crux-devel02:03
*** c0x has quit IRC02:05
jaegerI'll bootstrap and upload a test ISO tomorrow for whoever wants to run any tests02:57
*** `c0x has joined #crux-devel07:28
*** c0x` has quit IRC07:31
*** Romster has joined #crux-devel08:22
juejaeger: good news :)09:29
juefrinnst: thanks, they did a small code fix,
Romsterwhy don't they bump the patch version on that.10:11
frinnstlike sudo? :D10:14
frinnstrelease madness10:14
Romsterya lol11:05
Romsterat least you know when there is a new release with sudo11:12
prologicI prefer point release too over patches any day11:26
*** pitillo has quit IRC12:47
*** pitillo has joined #crux-devel12:49
frinnstand *fuck* github-only projects12:59
frinnst"releases are deprecated" crap12:59
Romsterthey expect us yo git clone to get the source now?13:05
Romstercan't see why they can't do foo-tag.tar.xz13:06
prologicthis will brighten your mood :)13:06
Romsterprologic, i just opened that twice now i was looking in your channel and already opened it13:07
prologicgood :)13:08
Romsterllvm why so long...13:25
frinnsti dont think you can host tarballs at github anymore13:28
Romsteruh oh13:36
jaegerinterestingly I'm able to boot a 3.1 ISO in virtualbox with UEFI if I use the kernel's builtin EFI stub support but the builtin command line portion doesn't work to load the initramfs13:48
jaegerso have to pass it the initrd via an EFI shell or something13:48
jaegerso dumb13:48
jaegerAnyway, uploaded a 3.1 test here for testing:
frinnstwtf, "sda1: WRITE SAME failed. Manually zeroing."14:14
frinnstafter partitioning (gpt) and mkfs.ext4 && mount14:14
frinnstsome new ext4 magick?14:14
jaegeryeah, no idea what causes that but see it frequently in VMs. some missing disk support or something, maybe?14:14
frinnstall devices were found afaik14:16
frinnstseems like a mptspi thingy14:17
jaegerlooks like a SCSI command that allows same blocks to be cloned in a range14:19
jaegerprobably just doesn't work with the virtual controllers14:19
jaegerhence the "manually zeroing" bit14:19
frinnstyeah, just surprised. never seen it before14:19
frinnstdoh, only two cores and too lazy to deselect kernel drivers14:27
frinnstthis might take a while D;14:27
frinnstwell, it boots :)14:39
frinnstenp2s0 D;14:39
jaegerthat's the new udev naming scheme. I kinda like it, they don't ever change14:40
jaegerYou can still override it if you prefer14:41
frinnstyeah, but i bet lots of people will be surprised and start debugging why their network driver wont work14:42
frinnstmaybe include 75-persistent-net-generator.rules ?14:42
jaegerI'd leave it out, myself, and make people be a little aware of the new system... if that's not the majority opinion, though, I won't cry about it14:44
frinnstim just worried about people in #crux whining "help, my network doesnt work" or whatever ;)14:45
frinnstrc.conf/net will probably need some tweaks either way14:45
frinnstdhcpcd is smart enough to find the nic, thats good14:46
jaegerI was thinking about that and considering that maybe setting up dhcp would be a better default in rc.d/net than manual config these days14:48
jaegerespecially since no interface is needed14:48
frinnstthe default config for rc.d/eth has always bothered me ;)14:58
jaegerI think it's just been carried forward as-is for a long time, we just updated the tools that were used14:59
jaegerip vs. ifconfig, etc.14:59
Romsterfinally got mesa3d-32 to compile.15:52
Romstertoo tired to commit it tonight i'll do that after work.15:52
Romsterthanks to jaegers work on llvm-32 i had todo a few tweaks but it works now.15:53
*** sepen has joined #crux-devel21:42
sepenis some 3.1 pre iso avail?21:43
sepenmy laptop at office runs win8 and I'm wondering about to install crux under virtualbox21:44
jaegerjust uploaded today. UEFI booting won't work with it but everything else should21:45
*** sepen has quit IRC21:59
*** sepen has joined #crux-devel22:22
sepenjaeger: just I saw some mirrors having 3.1 dir22:22
jaegersepen: I don't understand what you mean, sorry22:27
sepennah, I saw here a 3.1 directory
sepenbut seems only there are ports ;D22:29
nomiusI think those are just symlinks to the 3.0 ports22:30
jaegerthe 3.1 rsync repos do exist, aside from contrib22:30
jaeger(unless someone created contrib in the last few days while I was gone)22:30
jaegerjue: any comments/thoughts on the licensing question in #crux? About iso.git, prt-utils, etc.22:43
*** sepen has quit IRC23:00
*** sepen has joined #crux-devel23:59

Generated by 2.11.0 by Marius Gedminas - find it at!